摘要
设X_(1N),…,X_(NN)是相互独立的随机变量,它们的分布函数均连续,N=1,2,…。简单线性秩统计量的形状为其中C_(1N),…,C_(NN)是回归常数;a_N(1),…,a_N(N)是计分值;R_(iN)是X_(iN)在X_(1N),…,X(NN)中的秩。在一定的条件下,本文证明了S_N的大偏差概率的一致收敛区间为[0,o(N^(1/6-η))],其中η∈[0,1/6)。
Let X_(1N), …, X_(NN) be independent random variables having continuous distri-butions functions, N=1, 2, …. Simple linear rank statistics are the form of where c_(1N), …, C_(NN) are the regression constants, a_N(1), …, a_N(N) are the scores and R_(1N) is the rank of X_(1N) among X_(1N), …, X_(NN). Under certain conditions, it has been proved that the uniform convergence intervals of probabilities of large deviations for S_N are in this paper, where.
出处
《同济大学学报(自然科学版)》
EI
CAS
CSCD
1989年第2期217-227,共11页
Journal of Tongji University:Natural Science
关键词
秩统计量
大偏差
收敛
概率
Order statistics
Rank statistics
Large deviation-Probability
Uniform convergence-Interval